MN1–Fli1 oncofusion transforms murine hematopoietic progenitor cells into acute megakaryoblastic leukemia cells

Long-term outcome of acute megakaryoblastic leukemia (AMKL) patients without Down's syndrome remains poor. Founding mutations and chimeric oncogenes characterize various AMKL subtypes. However, for around one third of all cases the underlying mechanisms of AMKL leukemogenesis are still largely unknown. Recently, an in-frame fusion of meningeoma 1–friend leukemia virus integration 1 (MN1–Fli1) gene was detected in a child with AMKL. We intended to investigate the potential role of this oncofusion in leukemogenesis of acute myeloid leukemia. Strikingly, expression of MN1–Fli1 in murine hematopoietic progenitor cells was sufficient to induce leukemic transformation generating immature myeloid cells with cytomorphology and expression of surface markers typical for AMKL. Systematic structure function analyses revealed FLS and 3′ETS domains of Fli1 as decisive domains for the AMKL phenotype. Our data highlight an important role of MN1–Fli1 in AMKL leukemogenesis and provide a basis for research assessing the value of this oncofusion as a future diagnostic marker and/or therapeutic target in AMKL patients.


INTRODUCTION
Acute megakaryoblastic leukemia accounts for up to 2% of acute myeloid leukemia (AML) in adults. Distinct AMKL subtypes are characterized by founding mutations and chimeric oncogenes. [1][2][3] Prognosis of pediatric AMKL patients with Down's syndrome (DS) is favorable. However, very little is known about leukemogenesis of non-DS AMKL, and the outcome of these patients remains poor. Transcriptome sequencing of AML cells from pediatric AMKL patients revealed that AMKL is characterized by chimeric transcripts. 2 Various fusion proteins have been described in AMKL. 2,4,5 Recently, in one case of pediatric AMKL a novel in-frame fusion of meningeoma 1-friend leukemia virus integration 1 (MN1-Fli1) gene was identified. 2 In adult AML overexpression of both, MN1 and Fli1 is frequently found and confers a poor prognosis to the patients. [6][7][8] Moreover, Fli1 is a known key transcription factor in megakaryopoiesis. 9 Thus, we intended to investigate the role of MN1-Fli1 in adult AMKL development.

RESULTS AND DISCUSSION
To determine the role of MN1-Fli1 in AMKL, we retrovirally expressed MN1, MN1-Fli1 or an empty vector control in primary murine c-kit + hematopoietic progenitor cells (HPCs) derived from the bone marrow of C57Bl/6 mice and assessed their effect on in vitro colony replating capacity as a surrogate measure of selfrenewal. Expression levels of MN1 and MN1-Fli1, respectively, were similar as confirmed via quantitative reverse transcription-PCR (Supplementary Figure 1). In contrast to cells transduced with vector control, those transformed by MN1 could be serially replated to form compact colonies beyond the third round of replating (Figures 1a and b; and data not shown) generating myeloid leukemia cells as shown previously. 10 Interestingly, MN1-Fli1-transduced cells were able to form similar numbers of compact colonies resembling to MN1-derived colonies beyond the third round of replating (Figures 1a and b; and data not shown). However, May-Giemsa staining of cytospins of MN1 and MN1-Fli1 third round cells revealed strong differences in terms of morphology and expression of cell surface markers. Although MN1-transduced cells displayed their known myeloid blast morphology (Figure 1c), MN1-Fli1-transformed cells showed the phenotype of immature megakarypoietic cells, and mostly megakaryoblasts ( Figure 1c). Moreover, flow cytometry analysis of these cells showed significant differences in terms of cell surface marker expression. The large majority of MN1-Fli1transduced cells was positive for both CD41 and CD61, two characteristic AMKL markers, whereas double positivity for these markers was largely absent in MN1-transduced cells ( Figure 1d). As a next step, we performed structure function analyses to identify potential domains mediating AMKL phenotype to HPCs transformed by MN1-Fli1. The MN1-Fli1 oncoprotein is the in-frame fusion product of N-terminal MN1 sequence, a transcriptional coactivator, fused to C-terminal Fli1 sequence, an E26-transformation-specific (ETS) transcription factor. MN1-Fli1 contains almost the entire MN1 coding sequence at the N-terminal part and 470% of the Fli1 gene at the C terminus. 2 (Figures 2d and e).
Screening peripheral blood and bone marrow samples of 33 adult AMKL patients via PCR, we could not detect any patient with expression of MN1-Fli1 (data not shown). Thus, to finally determine the frequency and age distribution of this oncofusion in AMKL, analysis of a larger patient cohort including pediatric patients will be required.
The Fli1 gene encodes a member of the ETS family of transcription factors, which have an important role in hematopoiesis including the regulation of hematopoietic stem/progenitor cell maintenance and differentiation. 11,12 Moreover, aberrant expression of Fli1 is involved in myeloid leukemias and various types of human cancers including Ewing sarcoma with balanced chromosomal translocations generating the EWS-Fli1 oncofusion. 12 These studies implicate Fli1 and its signaling pathways as candidates for future targeted cancer therapies. 12,13 Fli1 binds to DNA in a sequence-specific way. Both, the FLS domain as well as the 3ʹETS domain of Fli1 contribute to DNA binding suggesting that MN1-Fli1-mediated leukemic transformation generates an AMKL phenotype via abnormal transcriptional activation of Fli1. We speculate that Fli1 fused to MN1 is redirected toward specific target gene promoters via FLS and 3ʹETS domain dependent DNA binding determining the AMKL phenotype.
Our study underscores a potential key role for the MN1-Fli1 oncofusion in AMKL leukemogenesis. Future studies may determine a potential diagnostic and/or prognostic role of MN1-Fli1 expression in AML as well as investigate its value as a putative marker for remission control in a subgroup of AMKL patients. Further investigation is needed to identify the major Diego, CA, USA) were used. Protocols and reagents used were as previously described. 14 Histographs represent staining obtained with antibodies specific for murine CD41 and CD61. Error bars represent s.d. from at least three independent experiments. Two-tailed Student's t-test was used to determine statistical significance (**P o0.05).
target genes of MN1-Fli1. Targeting either Fli1 directly or its key target genes may enable development of future therapeutic strategies for non-DS AMKL patients representing the poor prognosis subgroup of AMKL.  Histographs represent staining performed with antibodies specific for murine CD41 and CD61. Error bars represent s.d. from at least three independent experiments. One-way analysis of variance (ANOVA) was used to determine statistical significance (***P o0.0001).